Search results

The purpose of this research is to achieve multi-task autonomous driving by adjusting the network architecture of the model. Meanwhile, after achieving multi-task autonomous driving, the authors found that the trained neural network model performs poorly in untrained scenarios. Therefore, the authors proposed to improve the transfer efficiency of the model for new scenarios through transfer learning.

First, the authors achieved multi-task autonomous driving by training a model combining convolutional neural network and different structured long short-term memory (LSTM) layers. Second, the authors achieved fast transfer of neural network models in new scenarios by cross-model transfer learning. Finally, the authors combined data collection and data labeling to improve the efficiency of deep learning. Furthermore, the authors verified that the model has good robustness through light and shadow test.

This research achieved road tracking, real-time acceleration–deceleration, obstacle avoidance and left/right sign recognition. The model proposed by the authors (UniBiCLSTM) outperforms the existing models tested with model cars in terms of autonomous driving performance. Furthermore, the CMTL-UniBiCL-RL model trained by the authors through cross-model transfer learning improves the efficiency of model adaptation to new scenarios. Meanwhile, this research proposed an automatic data annotation method, which can save 1/4 of the time for deep learning.

This research provided novel solutions in the achievement of multi-task autonomous driving and neural network model scenario for transfer learning. The experiment was achieved on a single camera with an embedded chip and a scale model car, which is expected to simplify the hardware for autonomous driving.

With the rapid growth of the domestic lending industry, assessing whether the borrower of each loan is at risk of default is a pressing issue for financial institutions. Although there are some models that can handle such problems well, there are still some shortcomings in some aspects. The purpose of this paper is to improve the accuracy of credit assessment models.

In this paper, three different stages are used to improve the classification performance of LSTM, so that financial institutions can more accurately identify borrowers at risk of default. The first approach is to use the K-Means-SMOTE algorithm to eliminate the imbalance within the class. In the second step, ResNet is used for feature extraction, and then two-layer LSTM is used for learning to strengthen the ability of neural networks to mine and utilize deep information. Finally, the model performance is improved by using the IDWPSO algorithm for optimization when debugging the neural network.

On two unbalanced datasets (category ratios of 700:1 and 3:1 respectively), the multi-stage improved model was compared with ten other models using accuracy, precision, specificity, recall, G-measure, F-measure and the nonparametric Wilcoxon test. It was demonstrated that the multi-stage improved model showed a more significant advantage in evaluating the imbalanced credit dataset.

In this paper, the parameters of the ResNet-LSTM hybrid neural network, which can fully mine and utilize the deep information, are tuned by an innovative intelligent optimization algorithm to strengthen the classification performance of the model.

Existing studies have paid less attention to the impact of knowledge accumulation on digital transformation and its boundary conditions. Hence, this study aims to investigate the effects of ambidextrous knowledge accumulation on manufacturing digital transformation under the moderation of dynamic capability.

This study divides knowledge accumulation into exploratory and exploitative knowledge accumulation and divides dynamic capability into alliance management capability and new product development capability. To clarify the relationship among ambidextrous knowledge accumulation, dynamic capability and manufacturing digital transformation, the authors collect data from 421 Chinese listed manufacturing enterprises from 2016 to 2020 and perform analysis by multiple hierarchical regression method, heterogeneity test and robustness analysis.

The empirical results show that both exploratory and exploitative knowledge accumulation can significantly promote manufacturing digital transformation. Keeping ambidextrous knowledge accumulation in parallel is more conducive than keeping single-dimensional knowledge accumulation. Besides, dynamic capability positively moderates the relationship between ambidextrous knowledge accumulation and manufacturing digital transformation. Moreover, the heterogeneity test shows that the impact of ambidextrous knowledge accumulation and dynamic capabilities on manufacturing digital transformation varies widely across different industry segments or different regions.

First, this paper shifts attention to the role of ambidextrous knowledge accumulation in manufacturing digital transformation and expands the connotation and extension of knowledge accumulation. Second, this study reveals that dynamic capability is a vital driver of digital transformation, which corroborates the previous findings of dynamic capability as an important driver and contributes to enriching the knowledge management literature. Third, this paper provides a comprehensive micro measurement of ambidextrous knowledge accumulation and digital transformation based on the development characteristics of the digital economy era, which provides a theoretical basis for subsequent research.

The COVID-19 pandemic has profoundly impacted small and medium-sized enterprises (SMEs), inherently vulnerable entities, prompting a pivotal question of how to enhance SMEs’ organizational resilience (OR) to withstand discontinuous crises. Although digital innovation (DI) is widely acknowledged as a critical antecedent to OR, limited studies have analyzed the configurational effects of DI on OR, particularly stage-based analysis.

Underpinned by the dynamic capabilities view, this study introduces a multi-stage dynamic capabilities framework for OR. Employing Latent Dirichlet Allocation (LDA), digital product innovation (DPI), digital services innovation (DSI) and digital process innovation (DCI) are further deconstructed into six dimensions. Furthermore, we utilized fuzzy-set qualitative comparative analysis (fsQCA) to explore the configuration effects of six DI on OR at different stages, using data from 94 Chinese SMEs.

First, OR improvement hinges not on a singular DI but on the interactions among various DIs. Second, multiple equivalent configurations emerge at different stages. Before the crisis, absorptive capability primarily advanced through iterative DPI and predictive DSI. During the crisis, response capability is principally augmented by the iterative DPI, distributed DCI, and integrated DCI. After the crisis, recovery capability is predominantly fortified by the iterative DPI, expanded DPI and experiential DSI. Third, iterative DPI consistently assumes a supportive role in fortifying OR.

This study contributes to the extant literature on DI and OR, offering practical guidance for SMEs to systematically enhance OR by configuring DI across distinct stages.

All entrepreneurs interact simultaneously with multiple entrepreneurial contexts throughout their entrepreneurial journey. This conceptual paper has two central aims: (1) it synthesises the current literature on gender and entrepreneurship, and (2) it increases our understanding of how gender norms, contextual embeddedness and (in)equality mechanisms interact within contexts. Illustrative contexts that are discussed include entrepreneurship education, business networks and finance.

This conceptual paper draws upon extant literature to develop its proposed conceptual framework. It provides suggestions for systemic policy interventions as well as pointing to promising paths for future research.

A literature-generated conceptual framework is developed to explain and address the systemic barriers faced by opportunity-driven women as they engage in entrepreneurial contexts. This conceptual framework visualises the interplay between gender norms, contextual embeddedness and inequality mechanisms to explain systemic disparities. An extra dimension is integrated in the framework to account for the power of agency within women and with others, whereby agency, either individually or collectively, may disrupt and subvert the current interplay with inequality mechanisms.

This work advances understanding of the underrepresentation of women entrepreneurs. The paper offers a conceptual framework that provides policymakers with a useful tool to understand how to intervene and increase contextual embeddedness for all entrepreneurs. Additionally, this paper suggests moving beyond “fixing” women entrepreneurs and points towards disrupting systemic disparities to accomplish this contextual embeddedness for all entrepreneurs. By doing so, this research adds to academic knowledge on the construction and reconstruction of gender in the field of entrepreneurship.

The demand for halal food products is increasing globally. However, fraudulent activities in halal products and certification are also rising. One strategy to ensure halal integrity in the food supply chain is applying halal blockchain technology. However, to date, a few studies have assessed the factors and variables that facilitate or hinder the adoption of this technology. Thus, this study aims to assess the significant factors and variables affecting the adoption of halal blockchain technology.

A Delphi-based approach, using semi-structured interviews, was conducted with three food companies (chicken slaughterhouses, milk processing plants and frozen food companies). The cognitive best–worst method determines the significant factors and variables to prioritise halal blockchain adoption decisions.

The results showed that the most significant factors were coercive pressure and halal strategy. Nineteen variables were identified to establish a valid hierarchical structure for halal blockchain adoption in the Indonesian food industry. The five significant variables assessed through the best–worst method were demand, regulator, supply side, sustainability of the company’s existence and main customers.

The proposed halal blockchain decision structure can assist food companies in deciding whether to adopt the technology.

This study proposes 19 variables that establish a valid hierarchical structure of halal blockchain adoption for the Indonesian food industry.

The aim of this study is to discern the role of digital finance in driving rural industrial integration and revitalization. Specifically, it intends to shed light on how the deep development of digital finance can contribute to the optimization and transformation of the rural industrial structure. The research further explores the particular effects of this financial transformation in the central and western regions of China.

This research studies the influence of digital finance on rural industrial integration across 30 Chinese provinces from 2011 to 2020. Utilizing the entropy weight method, a comprehensive evaluation index system is established to gauge the level of rural industrial integration. A two-way fixed effects model, intermediary effect model, and threshold effect model are employed to decipher the relationship between digital finance and rural industrial integration.

Findings reveal a positive relationship between digital finance and rural industrial integration. A single threshold feature was identified: beyond a traditional finance development level, the marginal effect of digital finance on rural industrial integration increases. These effects are more noticeable in central and western regions.

Empirical outcomes contribute to policy discourse on rural digital finance, assisting policymakers in crafting effective strategies. Understanding the threshold of traditional finance development provides a new perspective on the potential of digital finance to drive rural industrial integration.

In the context of Industry 4.0, intelligent construction technologies (ICT) represented by information technology and networking will undoubtedly provide new impetus to the development of the prefabricated building supply chain (PBSC), but they will also bring various potential risks. So far, there is a large lack of research on the comprehensive consideration of the risks associated with the intelligent transformation of PBSC based on the information sharing perspective, and the critical risks and interactions are still unclear, making it difficult to identify efficient risk mitigation strategies. Therefore, this paper aims to reveal the interactions between stakeholders and clarify the critical risk nodes and interactions in information sharing of PBSC (IS-PBSC), and propose targeted risk mitigation strategies.

Firstly, this paper creatively delineates the risks and critical stakeholders of IS-PBSC. Secondly, Data is collected through questionnaires to understand the degree of risks impact. Thirdly, with the help of NetMiner 4 software, social network analysis is conducted and IS-PBSC risk network is established to reveal critical risk nodes and interactions. Finally, further targeted discussion of critical risk nodes, the effectiveness and reasonableness of the risk mitigation strategies are proposed and verified through NetMiner 4 software simulation.

The results show that the critical risks cover the entire process of information sharing, with the lack of information management norms and other information assurance-related risks accounting for the largest proportion. In addition, the government dominates in risk control, followed by other stakeholders. The implementation of risk mitigation strategies is effective, with the overall network density reduced by 41.15% and network cohesion reduced by 24%.

In the context of Industry 4.0, ICT represented by information technology and networking will undoubtedly provide new impetus to the development of the PBSC, but they will also bring various potential risks. So far, there is a large lack of research on the comprehensive consideration of the risks associated with the intelligent transformation of PBSC based on the information sharing perspective, and the critical risks and interactions are still unclear, making it difficult to identify efficient risk mitigation strategies.

Based on the results of risk network visualization analysis, this paper proposes an ICT-based IS-PBSC mechanism that promotes the development of the integration of ICT and PBSC while safeguarding the benefits of various stakeholders.

The potential of vehicle ad hoc networks (VANETs) to improve driver and passenger safety and security has made them a hot topic in the field of intelligent transportation systems (ITSs). VANETs have different characteristics and system architectures from mobile ad hoc networks (MANETs), with a primary focus on vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication. But protecting VANETs from malicious assaults is crucial because they can undermine network security and safety.

The black hole attack is a well-known danger to VANETs. It occurs when a hostile node introduces phony routing tables into the network, potentially damaging it and interfering with communication. A safe ad hoc on-demand distance vector (AODV) routing protocol has been created in response to this issue. By adding cryptographic features for source and target node verification to the route request (RREQ) and route reply (RREP) packets, this protocol improves upon the original AODV routing system.

Through the use of cryptographic-based encryption and decryption techniques, the suggested method fortifies the VANET connection. In addition, other network metrics are taken into account to assess the effectiveness of the secure AODV routing protocol under black hole attacks, including packet loss, end-to-end latency, packet delivery ratio (PDR) and routing request overhead. Results from simulations using an NS-2.33 simulator show how well the suggested fix works to enhance system performance and lessen the effects of black hole assaults on VANETs.

All things considered, the safe AODV routing protocol provides a strong method for improving security and dependability in VANET systems, protecting against malevolent attacks and guaranteeing smooth communication between cars and infrastructure.

In order to meet the different quality of service (QoS) requirements of vehicle-to-infrastructure (V2I) and multiple vehicle-to-vehicle (V2V) links in vehicle networks, an efficient V2V spectrum access mechanism is proposed in this paper.

A long-short-term-memory-based multi-agent hybrid proximal policy optimization (LSTM-H-PPO) algorithm is proposed, through which the distributed spectrum access and continuous power control of V2V link are realized.

Simulation results show that compared with the baseline algorithm, the proposed algorithm has significant advantages in terms of total system capacity, payload delivery success rate of V2V link and convergence speed.

The LSTM layer uses the time sequence information to estimate the accurate system state, which ensures the choice of V2V spectrum access based on local observation effective. The hybrid PPO framework shares training parameters among agents which speeds up the entire training process. The proposed algorithm adopts the mode of centralized training and distributed execution, so that the agent can achieve the optimal spectrum access based on local observation information with less signaling overhead.